Mocking stdin, stderr and stdout for python unittest
By Sophie Au
05 November 2018
This is a crosspost from Sophie's Blog.
Working on my latest project, todoster, forced me to learn all about input/output mocking using python's built-in
Mocking Output to stdout or stderr
The way I've been using the mock output or error stream is by getting its contents as a string. For this, there is a very simple method:
With this string you can do assertions as normal. However, since
mock_output aren't actually strings but streams you need to be very careful when trying to use the same stream a second time. The streams don't get cleared automatically so you need to either re-declare the mocks or make sure they're manually cleared out before every re-use.
If you want to have one mock for one test case (class) just make sure you're resetting
mock_err after every test like so:
import patch from unittest.mock class TestClass mock_out = patch('sys.stdout', new_callable=StringIO) mock_err = patch('sys.stderr', new_callable=io.StringIO) def tearDown(self): self.mock_out.truncate(0) self.mock_out.seek(0) self.mock_err.truncate(0) self.mock_err.seek(0) # ...
On the other hand, you can mock on a test by test basis by using decorators like so:
from unittest.mock import patch class TestClass(unittest.TestCase): #... @patch('sys.stderr', new_callable=io.StringIO) @patch('sys.stdout', new_callable=io.StringIO) test_small_functionality(self, mock_out, mock_err) # test something
Now you don't have to worry about clearing out the mocks after every test but it does come with having to add a new decorator for every test method that needs the mock.
If you want to mock stdin, this is the simplest solution I could find. It assumes that the code under test calls input only once though.
@patch('builtins.input', side_effect=['the input you want to test']) def test_delete_project_say_no(self, _): # test something